Current debate on the use of antibiotic prophylaxis for caesarean section
Dr RF Lamont and Dr R Romero, Perinatology Research Branch, NICHD, NIH, DHHS, Wayne State University/Hutzel Women’s Hospital, 3990 John R, Box 4, Detroit, MI 48201, USA. Emails email@example.com and firstname.lastname@example.org
Please cite this paper as: Lamont R, Sobel J, Kusanovic J, Vaisbuch E, Mazaki-Tovi S, Kim S, Uldbjerg N, Romero R. Current debate on the use of antibiotic prophylaxis for caesarean section. BJOG 2011;118:193–201.
Caesarean delivery is frequently complicated by surgical site infections, endometritis and urinary tract infection. Most surgical site infections occur after discharge from the hospital, and are increasingly being used as performance indicators. Worldwide, the rate of caesarean delivery is increasing. Evidence-based guidelines recommended the use of prophylactic antibiotics before surgical incision. An exception is made for caesarean delivery, where narrow-range antibiotics are administered after umbilical cord clamping because of putative neonatal benefit. However, recent evidence supports the use of pre-incision, broad-spectrum antibiotics, which result in a lower rate of maternal morbidity with no disadvantage to the neonate.
Before the mid-nineteenth century, surgical procedures commonly resulted in postoperative sepsis and death. In the 1860s, when Joseph Lister introduced the principles of anti-sepsis (Table 1), the incidence of postoperative infectious morbidity and mortality fell markedly from 50 to 15%. In the 1960s, using an animal model, Burke1 demonstrated that if antibiotics were given before wound contamination, the rate of infection decreased. Following caesarean delivery (CD), maternal mortality and morbidity may result from a number of infections including endometritis, urinary tract infection (UTI) and surgical site infection (SSI),2 which if deep rather than superficial, increase hospital stay and cost per case.3–11
Table 1. Innovations based upon antiseptic principles in the practice of surgery attributed to Joseph Lister which reduced the rate of post-surgical infectious mortality and morbidity1
|Carbolic acid (phenol) wound dressings|
|Absorbable (catgut) sutures soaked in carbolic acid|
|Surgical hand washing with carbolic acid|
|Sterilisation of surgical instruments with carbolic acid|
|Use of surgical drainage tubes|
|Use of gloves masks and gowns|
Following elective surgery, wound infection in patients who receive perioperative antibiotics (within 3 hours following skin incision) occurs in 1.4% compared to 0.6% in those who receive antibiotics within the 2 hours before skin incision.12,13 Prophylactic antibiotics reduce the incidence of SSIs14–16 especially those used before incision rather than during or after the procedure.14,15,17,18 Single dose antibiotic prophylaxis is well-established for abdominal and vaginal hysterectomy,19,20 and cumulative meta-analysis data indicate that if the various studies had been pooled at an earlier date, the use of controls in subsequent trials who received no treatment would have been unnecessary.21 By the 1990s, 27 million surgical procedures were performed annually in the USA,22 and SSIs accounted for approximately 15% of all nosocomial infections.23 Prophylactic antibiotics, have been shown to reduce the rate of SSIs after elective surgical procedures24 and are increasingly used as performance indicators.25 Postpartum SSIs, especially those following CD, are more common than those following other surgical procedures.26
Other infections occurring after vaginal or abdominal birth
For most pregnant women, SSIs are not life threatening, but they have important implications on the length of hospital stay, hospital costs and social implications for the parents and the newborn.27–29 In the USA, infection still accounts for a disproportionate contribution to pregnancy-related mortality.2 Caesarean delivery is the single most important factor associated with postpartum infection,30,31 and carries a five-fold to 20-fold increased risk of infection compared with vaginal delivery,31 though this varies according to definition, classification and duration of observation.30,32–43 The rate of infection following CD is 1.1–25% compared with 0.2–5.5% following vaginal birth,27,30,31,33–35,38,42,44–48 and the rates of infection after emergency and elective CD vary between 7.5 and 29.8% and between 5.5 and 17.3%, respectively.28,46,49–55 The most common infection-related complication following CD is endometritis, 16,56 which can be reduced by 50% with the use of prophylactic antibiotics.16,57–59 The prevalence of postpartum UTI and wound infection (WI) varies geographically,34,42,44,46,49,55,60–63 but with or without endometritis, WI complicates more than 10% of CDs despite recommended antibiotic prophylaxis,16 and is 50% higher in emergency compared with elective CD.64 Infection within 30 days following CD occurs in 7.6% of women compared with 1.6% following vaginal birth,64 but these numbers may be misleading because CD may be elective or emergency, each of which will carry a different risk of infection. Up to 80% of infections occur after discharge from the hospital;15,34,38,46,49,54,55,60,61,64,65 therefore, post-CD infection rates may be underestimated if based on hospital discharge records.
The risk factors associated with infection following CD are extensive.31,60,66–71 A major risk factor for postoperative infection is emergency CD (compared with elective).39,45,47,72,73 In addition, high maternal body mass index, failure to use surgical drainage with subcutaneous tissue thickness of ≥3 cm, as well as prolonged operating time and poor surgical technique are established major risk factors for post-CD infections.28,31,37,42,43,45,54,74–77 The most important risk factor is the host herself and a risk index category has been derived,78 validated79 and modified26 by the American College of Surgeons which comprises three variables: 1) the American Society of Anesthesiologists’ preoperative assessment score (capturing variables such as diabetes, obesity, pre-eclampsia); 2) length of operation; and 3) surgical wound classification (whether clean or contaminated, reflecting risks such as prolonged labour or rupture of the membranes).
Increasing caesarean delivery rate
Despite the World Health Organization’s estimate that caesarean section rates should not be >15%,80 in the developed world, CD rates are already above 20%.54,64,81 The CD rate increased by 50% in the USA between 1996 and 2006,82 and caesarean section is now the most commonly performed major surgical procedure,83 comprising 31% of births or 1.3 million procedures annually,65 which could reach 50% (or two million CDs each year) by 2020.84 Part of the reason for the increase in CD rates is the increased use of primary CD on maternal request64,85–88. In a survey of US obstetricians,89 53% of respondents confirmed that they had performed such procedures, 58% recorded that the rate had increased in the last year and 41% said that they routinely discussed the topic with women.90
Are prophylactic antibiotics effective in reducing infectious morbidity after CD?
Antibiotic prophylaxis for women undergoing CD has been proven to be beneficial in decreasing post-CD infectious morbidity both in women at high-risk (in labour after membrane rupture), or low-risk (non-labouring with intact membranes)16,91–94 (Table 2). A single dose of antibiotics is as effective as multiple doses given perioperatively,95–98 and the routine use of prophylactic antibiotics reduces the risk of infection by more than 50% from a baseline as high as 20–50%.99,100 Although antibiotic prophylaxis for elective CD has been shown to be cost effective,35,101 there has been reluctance to implement the recommendations,102,103 as well as inconsistency in their implementation.104,105 Several questions have been raised including the optimal indication, drug of choice and drug regimen,106–109 and whether prophylaxis should be given to all women or only those considered to be at high risk.35,110–112
Table 2. The effect of any prophylactic antibiotic on the rates of fever, wound infection, endometritis and urinary tract infection following caesarean delivery (adapted from Cochrane Systematic Review16)
|Fever||0.49 (0.32–0.75)||0.40 (0.31–0.51)|
|Wound infection||0.73 (0.53–0.99)||0.36 (0.26–0.51)|
|Endometritis||0.38 (0.22–0.64)||0.39 (0.34–0.46)|
|Urinary tract infection||0.57 (0.29–1.11)||0.43 (0.30–0.60)|
Current debate on the use of antibiotic prophylaxis for CD
Currently, the Cochrane Database of Systematic Reviews, the American College of Obstetricians and Gynecologists and the Centers for Disease Control16,91,113 recommend narrow-range first-generation cephalosporins, like cefazolin, to be administered after umbilical cord-clamping for prophylaxis against infection after CD. This is because they are considered equally effective and less costly than broad-spectrum antibiotics.91,114 However, despite the use of antibiotics, 10% of CD are still complicated by infection and 15% by fever.16 The administration of antibiotics is not intended to sterilise tissues, but to act as an adjunct to decrease the intra-operative microbial load to a level that can be managed by the host innate and adaptive immune responses.14,15,18,115 The goal of antibiotic therapy is to achieve sufficient tissue levels at the time of microbial contamination,1 and the optimal agent should be long-acting, inexpensive and have a low adverse effect profile.91 Ampicillin reaches Group B Streptococcus (GBS) bacteriocidal concentrations in cord blood within 5 minutes of administration to the mother,116 and cefazolin reaches the minimum inhibitory concentration for GBS in fetal blood within 30 minutes of administration.13 Concerns about antibiotic intervention using broad-spectrum antibiotics centre on infection with resistant organisms such as Clostridium difficile or methacillin-resistant Staphylococcus aureus, but this is unlikely with single-dose prophylaxis.15 In addition, the shorter hospital stay observed following antibiotic prophylaxis is not consistent with an increase in infections from resistant organisms.99
There is overwhelming evidence for the need and effectiveness of prophylactic antibiotics to prevent infection following CD, so the current debate focuses on the choice of antibiotic and the timing of administration. With respect to timing, the debate lies between pre-incision or after clamping of the umbilical cord and the choice of antibiotic lies between narrow-range and broad-spectrum. Both of these debates have been influenced by concerns that broad-spectrum antibiotics given before incision might mask neonatal infection or result in a neonatal infection in which no organism could be cultured. There are also concerns that the wrong choice of antibiotic may result in the neonate being exposed to resistant strains of bacteria,91,117 which might lead to a worse neonatal outcome118 or the need for expensive neonatal septic screens and infection work-ups.119 This is supported by an observed shift in early neonatal sepsis from GBS to Escherichia coli and other Gram-negative organisms, and a change in resistance patterns.91,118,120–122 This may affect early gut colonisation and, together with CD itself,123 has been implicated in early childhood asthma and allergy.124–131 If antibiotics for GBS prophylaxis are included, depending on the protocol used (screening-based or risk-based), a significant proportion of babies could be exposed to antibiotics during pregnancy with little known about the potential for long-term adverse effects.
Timing of administration of prophylactic antibiotics for CD
While regulatory agencies overwhelmingly advise that prophylactic antibiotics should be given before incision to prevent SSIs,14,15,18 an exception is made for CD, where the recommendation is to use these antibiotics after clamping of the umbilical cord.15,16 Recently, a systematic review of the literature has challenged this approach.84 A total of 277 potentially relevant studies were identified, from which two non-randomised trials,119,132 two retrospective-cohort studies24,133 and three randomised controlled trials (RCTs)134–136 were selected to produce a meta-analysis.137 The two non-randomised trials119,132 concluded that there was no benefit of pre-incision versus post-clamping antibiotic prophylaxis with respect to overall infection rate or endometritis. However, one of the studies was unblinded, with a small sample size,132 and the other was a secondary analysis of two trials,119 one of which used narrow-range, and the other, broad-spectrum antibiotics. Most meta-analyses or systematic reviews would exclude such studies. The same is true of the two retrospective cohort studies, both of which used narrow-range cephalosporins,24,133 yet found a significant reduction in overall infection rates. One showed a significant reduction in endometritis,24 and the other reported a significant reduction in WI133 when prophylactic antibiotics were given pre-incision rather than post-clamping. In the meta-analysis137 of the three RCTs134–136 with a sample size of 749, the use of pre-incision cefazolin was associated with a 50% reduction in overall infection rate, a 53% reduction in the rate of endometritis and a non-statistically significant 40% reduction in the rate of WI compared with post-clamping administration. Neonatal sepsis rates were comparable between the two study groups.
Subsequently, a retrospective cohort study of 1316 term singleton CDs at one institution reported on a policy change in timing of antibiotic prophylaxis from post-clamping to pre-incision, which resulted in a reduction of 60% in the rate of SSIs, a 50% reduction in the rate of endometritis and an 80% decrease in cellulitis.24 Another recent study supports these findings,138 though it was not an RCT. The observation arose from a change in policy from post-clamping administration of prophylactic antibiotics to pre-incision administration over two different time periods. With >4000 CDs in each group and using the same antibiotics throughout (cefazolin), there were no adverse neonatal effects. Post-clamping antibiotics (n = 4229) were associated with a 3.9% incidence of endometritis compared with 2.2% incidence for pre-incision cefazolin (n = 4781) (adjusted odds ratio 0.61; 95% CI 0.47–0.79), and post-clamping WI occurred in 3.6% compared with 2.5% following pre-incision antibiotics (adjusted odds ratio 0.70; 95% CI 0.55–0.90; P = 0.001 for the linear trend).138 In contrast, using cefazolin in women undergoing elective CD, pre-incision antibiotics did not significantly reduce overall infection or endometritis. Nevertheless, the use of pre-incision antibiotics was not associated with an increase in neonatal sepsis, sepsis work-up, and admission or length of stay in the Neonatal Intensive Care Unit.139
Choice of antibiotic and rationale for the use of broad-spectrum antibiotics
The main source of infection following CD is the lower genital tract51,140 particularly if the membranes are ruptured, but this still occurs with intact membranes, especially following preterm birth.141–143 The causative organisms are polymicrobial, particularly those responsible for bacterial vaginosis, such as Ureaplasma spp., Mycoplasma spp., anaerobes or Gardnerella vaginalis37,57,70,144–151 and these organisms are also commonly isolated from amniotic fluid and the chorioamnion at the time of CD.57,70,152–154 When these organisms are detected, there is a three- to eight-fold increased risk of endometritis or WI after CD,57,84,152–155 and bacterial vaginosis is associated with a six-fold increase in post-CD endometritis.70 Wound infection is susceptible to skin contaminants as well as to organisms responsible for bacterial vaginosis.144,147
The use of first-generation cephalosporins such as cefazolin91 provides antibiotic activity against species of Ureaplasma and Mycoplasma but may cause an increase in resistant organisms like anaerobes.146,156 Hence, there is a rationale for adding agents such as metronidazole, clindamycin or azithromycin to extend the cover. The broad-spectrum antibiotics that have been evaluated are mainly single-agent extended-range penicillins, or second- or third-generation cephalosporins (β-lactams), which show no advantage.114 However, four RCTs99,157–159 compared the use of narrow-range antibiotic prophylaxis (first-generation cephalosporin or ampicillin) with broad-spectrum regimens which comprised narrow-range antibiotics with the addition of agents from a different class of antibiotics such as gentamycin,158 metronidazole157,159 or azithromycin and doxycycline.99 Broad-spectrum antibiotics were associated with a statistically significant reduction in infection rates,99 endometritis99,157–159 and WI99 compared with narrow-range agents. Length of hospital stay was significantly shorter when broad-spectrum antibiotics were used.99,157,158 The leading option as a second-line broad-spectrum antibiotic for CD appears to be azithromycin, which has a longer half-life (68 hours), higher tissue concentrations, and lower transplacental passage than several other antibiotics commonly used for this indication.91 In addition, azithromycin is active against both aerobes and anaerobes, as well as Ureaplasma spp., resulting in significantly less endometritis and WI than in studies in which other antibiotics were used.99,157–161 Metronidazole is cheaper than azithromycin, but as 20% of preterm neonates may have Ureaplasma bacteraemia162 and it is suggested that neonatal Ureaplasma infection may be associated with bronchopulmonary dysplasia,163 azithromycin-based broad-spectrum prophylaxis may prevent neonatal sepsis and chronic lung disease, though this has not been tested. Additional support for the use of azithromycin-based broad-spectrum antibiotic prophylaxis for CD has been demonstrated in a series of studies with experience obtained through institutional surveillance.57,99,160,161
Experience obtained through institutional surveillance
The first of these studies included 575 women undergoing CD with intact membranes and no evidence of chorioamnionitis. Colonisation of the chorioamnion with Ureaplasma urealyticum, irrespective of the presence of other organisms, was associated with a three-fold increased risk of endometritis, which rose to an eight-fold increased risk if the women had gone into spontaneous labour.57 Subsequently, in an RCT of 597 women undergoing CD, using broad-spectrum antibiotics known to be active against U. urealyticum, the prevalence of endometritis, WI or either, was statistically significantly reduced compared with the use of cefotetan and placebo. Length of hospital stay overall and the rate of endometritis were also statistically significantly reduced.99 Finally, institutional surveillance over a 14-year period demonstrated that when comparing the time during which narrow-range antibiotics were used, with the interim period of trials during which broad-spectrum antibiotics were tested, the latter was associated with a finite reduction in post-CD infection rate.160,161 Endometritis rates fell from 23% with narrow-range, to 16% during the trial period, to 2.1% with routine use of broad-spectrum antibiotics.161 Wound infection showed the same trend. During the use of narrow-range antibiotics, the rate of WI fell from 3.1 to 2.4% during the trial period, and to 1.3% with the routine use of broad-spectrum antibiotics.160
Caesarean delivery is associated with a significantly higher postoperative infection (SSI, WI, UTI) rate than following vaginal birth and other surgical procedures. With the increase in CD rates worldwide, such post-CD infections are likely to become a significant health and economic burden. There is overwhelming evidence that antibiotic prophylaxis for CD is effective in preventing maternal infectious morbidity. However, concerns about neonatal infection have confined its use to narrow-range antibiotic administration after umbilical-cord clamping, instead of the regimen of pre-incision, broad-spectrum antibiotics that is used in non-pregnant patients undergoing major surgery. Recent evidence suggests that pre-incision broad-spectrum antibiotics are more effective in preventing post-CD infections than post-clamping narrow-range antibiotics, without prejudice to neonatal infectious morbidity. This strategy has been adopted by the American College of Obstetricians and Gynecologists and the American Academy of Pediatricians,164 though national guidelines have yet to change. Nevertheless, the combination of broad-spectrum/pre-incision antibiotic prophylaxis for CD versus narrow-range/post-clamping has not been tested and there is an urgent need for this definitive study to be performed. Such a study would have to address both maternal and neonatal infectious morbidity as well as long-term neonatal follow up. Variables such as surgical technique (suture material, use of surgical drainage),15,77 type of CD (elective versus emergency; primary versus repeat; with or without labour) and state of the chorioamniotic membranes would have to be addressed.
Disclosure of interests
Contribution to authorship
All the authors contributed to the production and editing of the manuscript.
Details of ethics approval
This research was supported in part by the Perinatology Research Branch, Division of Intramural Research, Eunice Kennedy Shriver National Institute of Child Health and Human Development, NIH, DHHS.